The traditional surrogate measures of safety (like number of vehicles in dilemma zone) fail to quantify the risk of crash. Traffic conflict has been proposed as an improved surrogate measure of safety for operations at isolated intersections to quantify the risk. This paper develops a dilemma zone hazard function estimating procedure to obtain the probability of traffic conflict occurring. This approach is an extension of the current approach of dilemma zone boundaries to determine the risk of traffic conflict for an individual vehicle in the case of a dilemma zone incursion. Field data collected from the intersection of SR37 and SR32 at Noblesville, Indiana is used to generate a binary choice model that best explains the underlying criteria for a driver’s decision at the onset of yellow. The probability of making an erroneous decision is used as the probability of traffic conflict (dilemma hazard function). An economic framework was developed to implement the dynamic of dilemma hazard function using existing controllers. Although the data are specific to one intersection, the procedures are readily transferable. This paper also demonstrates the potential of sensor providing richer data than an inductive loop detector can be used to further enhance the safety at signal operations.